RSSNet is a deep separation network that decomposes single-channel noisy Raman spectra into pure component spectra, outperforming sparse regression by over 4 dB on synthetic data and generalizing from synthetic training to real mineral powder mixtures.
An efficient encoder-decoder architecture with top-down attention for speech separation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Brain-Inspired Deep Separation Network for Single Channel Raman Spectra Unmixing
RSSNet is a deep separation network that decomposes single-channel noisy Raman spectra into pure component spectra, outperforming sparse regression by over 4 dB on synthetic data and generalizing from synthetic training to real mineral powder mixtures.